notesum.ai

Published at October 30

$\textbf{EMOS}$: $\textbf{E}$mbodiment-aware Heterogeneous $\textbf{M}$ulti-robot $\textbf{O}$perating $\textbf{S}$ystem with LLM Agents

cs.RO

cs.AI

cs.MA

I.2.7; I.2.8; I.2.9; I.2.10

Released Date: October 30, 2024

Authors: Junting Chen¹, Checheng Yu¹, Xunzhe Zhou², Tianqi Xu³, Yao Mu⁴, Mengkang Hu⁴, Wenqi Shao⁵, Yikai Wang⁶, Guohao Li⁷, Lin Shao¹

Aff.: ¹National University of Singapore; ²Fudan University; ³KAUST; ⁴The University of Hong Kong; ⁵Shanghai AI Laboratory; ⁶Tsinghua University; ⁷University of Oxford

Arxiv: http://arxiv.org/abs/2410.22662v1

Method	Succ. Rate $\uparrow$	Sub-goal Succ. Rate $\uparrow$	Token Usage $\downarrow$	Simulation Step $\downarrow$
EMOS (Ours)	37.82%	81.26%	80783	2358
w/o. Numerical	23.56%	71.04%	53201	2983
w/o. Robot resume	15.63%	65.27%	64600	3125
w/o. Discussion	15.23%	72.45%	36377	2332